[x264-devel] Re: x264: svn commit r240 (pengvado)

[Stephen Henry]

Hi,

I was having a look over the source code and have some questions
regarding how the multithreading has been implemented. Mainly, my
concern is with INTRA prediction.

It appears to me that the code has been modified so that the encoder
encodes 'i_threads' number of slices over a rectangular strip in the
frame. The encoding occurs simultaneously. How does the encoder enforce
INTRA prediction across slice boundaries? It looks to me that
'constrained intra prediction' isn't used and that for those macroblocks
neighboring different slices they use the original pixel values, opposed
to the reconstructed prediction values, for INTRA prediction. This would
mean that the prediction may be incorrectly reconstructed under certain
circumstances.

Regards,

Stephen Henry

-----Original Message-----
From: x264-devel-bounce at videolan.org
[mailto:x264-devel-bounce at videolan.org] On Behalf Of champ yen
Sent: 30 May 2005 05:36
To: x264-devel at videolan.org
Subject: [x264-devel] Re: x264: svn commit r240 (pengvado)

On 5/30/05, Subversion daemon <svn at videolan.org> wrote:
> r240 | pengvado | 2005-05-29 20:28:49 +0200 (Sun, 29 May 2005) | 2
lines
> Changed paths:
>    M /trunk/common/common.c
>    M /trunk/common/common.h
>    M /trunk/common/macroblock.c
>    M /trunk/common/macroblock.h
>    M /trunk/configure
>    M /trunk/encoder/analyse.c
>    M /trunk/encoder/cabac.c
>    M /trunk/encoder/encoder.c
>    M /trunk/x264.c
>    M /trunk/x264.h
> 
> multithreading (via slices)
> 
Well done! I've read the code. It's a better implementation.
I don't know how to contribute more to x264.
Anyway, I'm glad to do something more.

best regards
champ yen

-- 
This is the x264-devel mailing-list
To unsubscribe, go to: http://developers.videolan.org/lists.html



-- 
This is the x264-devel mailing-list
To unsubscribe, go to: http://developers.videolan.org/lists.html